When the chat says "this is the final version", that's the moment to bottle: the scenario opening, the differentiators, the 170-word cap — extracted at peak convergence.
The clearest distillation signal a conversation produces is the explicit declaration: "this is the final version." Everything before it was the path; everything in force at that moment is the asset. This extraction targets that moment — the template-language rejection, the scenario-opening correction, the must-include differentiators, the word cap — and bottles the state the declaration blessed. This setup loads an App Store description chat ending in exactly that declaration, extracted at the moment of peak convergence.
Catch the declaration
"This is the final version" is the extraction trigger — the state at that line is the asset.
Bottle everything in force
Rejections, corrections, constraints — all rules active at the declaration travel.
Revise from the bottle
Next quarter's update starts from the extracted prompt, not from archaeology.
The REQUIREMENTS block is ordered and labeled so later items override earlier ones where they conflict, mirroring how the conversation resolved with the last decision winning. Conversation-to-Prompt Builder bottles this; the QUALITY BAR also says to flag anything ambiguous or conflicting and ask, rather than silently pick one reading, when you rerun it in your assistant.
So rejected directions cannot be rediscovered. The AVOID section carries what was scrapped, like "ultimate companion" flagged as template language, and the WHY section notes the model cannot walk back paths you already rejected. It is kept separate from the active REQUIREMENTS, such as the under-170-words cap that the iteration actually earned.
Read it once. The USAGE NOTES report a Distillation confidence of High and state signals were extracted VERBATIM from the conversation, deduplicated but never paraphrased, then add to read it once before trusting it with a real run. The PRESERVED REQUIREMENTS counts help you check nothing important, like the standing rule, was dropped.
What a converged conversation contains: a rejected voice, an accepted framing, a hard constraint, a format decision, and a final-version moment — each mapped to its prompt section.
The result you finally liked lives inside the chat. Distill the iterations into one prompt: corrections become requirements, the accepted version becomes the quality bar.
In feedback threads the rejections carry the knowledge: the mocking empty-state enthusiasm, the passive phrasing — converted into the avoid-rules that define the style.
End a working session like a shift change, not an abandonment: state captured, decisions logged, next step named — ready for the next session to pick up.
Turns a reusable API request prompt into a clean variable set covering endpoint, method, auth, payload, response format, and error handling.
A reusable AI agent task template with variables for objective, context, available tools, constraints, success criteria, failure handling, and output format.
Distill an iterated conversation into one reusable prompt — corrections become requirements, rejections become an avoid-list.
After a dozen corrections, the AI finally gives you what you wanted — but the prompt that produces it is smeared across the whole chat. Here's how to distill it into one reusable prompt you can run again.